Neural mechanisms of face emotion processing in youths and adults with bipolar disorder

Abstract

Objectives: Little is known about potential differences in the pathophysiology of bipolar disorder (BD) across development. The present study aimed to characterize age-related neural mechanisms of BD.

Methods: Youths and adults with and without BD (N = 108, age range = 9.8-55.9 years) completed an emotional face labeling task during fMRI acquisition. We leveraged three different fMRI analytic tools to identify age-related neural mechanisms of BD, investigating (a) change in neural responses over the course of the task, (b) neural activation averaged across the entire task, and (c) amygdala functional connectivity.

Results: We found converging Age Group × Diagnosis patterns across all three analytic methods. Compared to healthy youths vs adults, youths vs adults with BD show an altered pattern in response to repeated presentation of emotional faces in medial prefrontal, amygdala, and temporoparietal regions, as well as amygdala-temporoparietal connectivity. Specifically, medial prefrontal and lingual activation decreases over the course of repeated emotional face presentations in healthy youths vs adults but increases in youths with BD compared to adults with BD. Moreover, youths vs adults with BD show less medial prefrontal activation and amygdala-temporoparietal junction connectivity averaged over the task, but this difference is not found for healthy youths vs adults.

Conclusion: Although longitudinal confirmation and replication will be necessary, these findings suggest that neural development may be aberrant in BD and that some neural mechanisms mediating BD may differ in adults vs children with the illness.

Trial registration: ClinicalTrials.gov NCT00025935 NCT00006177.

Keywords: bipolar disorder; development; fMRI; face emotion; pediatric.

Conflict of interest statement

Disclosures

The authors have no conflicts of interest to declare.

Published 2019. This article is a U.S. Government work and is in the public domain in the USA.

Figures

Figure 1.. Age Group x Diagnosis interaction predicts time course of activation.

Predicted slopes of activation change over time course of the first run of the task in the A) medial prefrontal cortex and B) lingual gyrus where an Age Group x Diagnosis interaction was evident. For this and all figures, brain images represent axial sections (left=right) with threshold set at whole-brain-corrected p<.05. Cluster intercepts and slopes were extracted and averaged for plots. Post-hoc comparisons indicate differences in slope in the following: HA>HY, BPY>BPA, BPY>HY, BPY>HA, BPY>0, HY<0 for (A); HA>HY, BPY>BPA, BPY>HY, HA>BPA, BPY>HA, HY<0 for (B), where HA=healthy adults, HY=healthy youths, BPA=bipolar adults, BPY=bipolar youths (all p<.05, corrected). Please see Supplemental Table 1 for Boxplots of the raw data of these clusters.

Figure 2.. Age Group x Diagnosis x Emotion and Age Group x Diagnosis interactions predict activation averaged over the course of the task.

A) Significant Age Group x Diagnosis x Emotion interaction in the lingual gyrus. B) Significant Age Group x Diagnosis interaction in the medial prefrontal cortex. *p<0.05, corrected.

Figure 3.. Age Group x Diagnosis predicts right amygdala connectivity with the bilateral temporo-parietal junction.

Graph displays significant Age Group x Diagnosis interaction in the left temporo-parietal junction. Patterns are similar in the right TPJ (illustrated in brain image; not shown in plot). *p<0.05, corrected.